This site uses cookies. By continuing to browse the site you are agreeing to our use of cookies.Find out more here


To discuss how you and your organisation can get more involved with The Work Foundation, please contact us.

Call 020 7976 3575 or email


Ian Brinkley
Economic Advisor
Ian Brinkley

Do we have an hourglass labour market?

Authors: Ian Brinkley

21 January 2015

 A recent paper published by Dr Holmes for the authoritative Centre for Skills, Knowledge and Organisational Performance (SKOPE) at Oxford University and featured in the Financial Times on 20 January suggests that the UK has had an hour-glass labour market, with growth at the top and bottom of the labour market, with jobs being lost from the middle over the period 1996 to 2008. This somewhat contradicts other analysis which suggests that over the past 15 years the net growth has been skewed towards high skill jobs, and the recent statement by the Prime Minister that most of the new jobs created since 2010 have also been high skill.

 One reason we may get different answers depends what we classify as a high, middle, and lower skill job. Many analysts including the Bank of England use a relatively simple classification based on the official statistics on occupations, which are themselves partly based on an assessment of relative skilled levels as well as what the jobs involve. The top three occupational categories in the official statistics – managers, professionals and associate professionals – as high skill; the next two categories of skilled manual and secretarial and administrative as mid skill; and the bottom four categories of care and leisure; sales related; semi-skilled operatives; and less skilled. The analysis by Dr Holmes and colleagues uses similar international occupational codes but groups them in a different way. The main difference is the inclusion of operatives as a middle skill group.

 It is not obvious one categorisation is better than the other – on the assumption that there is some relationship between pay and skills, operatives sit mid-way between the unambiguously lower skill groups and the unambiguously high skill group measured by hourly pay rates. However, looking at occupational change is not straightforward as jobs content change, new jobs are added, and old jobs drop out and as a result the way jobs are classified in the official statistics change over time. Indeed, some have concluded that the lack of clear-cut evidence on wage polarisation suggests that some of the middle skill jobs of the 1990s are not the same middle skill jobs of the current day, so that the “middle” has to some extent been recolonized by different occupations.

 We have readily available consistent UK statistics on the latest definitions for the period 2004 to 2014 . These show that between 2004 and 2014, the top three occupations increased by just over 2 million; the middle jobs using the SKOPE classification fell by just over 900,000; and less skilled jobs increased by 540,000. The common thread is that of employment polarisation, with a significant decline in middle jobs however defined. However, these figures suggest that in more recent periods the overall shape has moved away from the hourglass of the 1980s and 1990s to something more like a cocktail glass with net job growth concentrated in the higher skill categories.

 The OECD has recently produced an analysis covering a similar period to that in the SKOPE analysis, but has used a different definition entirely. The OECD classifies jobs as high, medium and low skill depending on level of educational attainment. It is not ideal, as educational attainment and skills are not the same thing. But it at least shows what has been happening to the jobs of those who hold a degree equivalent, those with secondary level education, and those with only basic schooling.

By this measure the biggest increase in employment has been for those with degree level education between 1995 and 2008 in the UK and across the EU15, the US, and Japan; followed by those with middle level education; and big falls in employment for those for lower levels of education. What these figures do not tell us is how much this increase is caused by the expansion of jobs for those with a university education and how much by an excess supply of graduates taking jobs that do not need a degree. My view is that it is much more of the former – the wage premia associated with degrees has not fallen in most OECD economies, including the UK– though that does not preclude some mis-match in the labour markets where graduates are currently seeking work.

 What the OECD analysis also shows is why these employment changes are taking place, looking at the impact of new technologies, trade and consumption. These effects are of course linked – for example, the direct effects of both trade and technology tend to depress employment in advanced economies but by boosting wages through productivity gains, driving down prices, opening up new markets, and creating new consumer demand they indirectly create more jobs than they destroy. The OECD provides estimates for the major countries and economic blocs and in the table below I have replicated those for the EU15 and the UK.
Employment cahnge bt education level across Europe

 The OECD concludes that skill-biased technological change is the main reason why employment for the less well educated is falling and that for the well-educated is growing. Consumption effects can mitigate the worst of these impacts and enhance employment growth for those with mid to high levels of education. Improving educational prospects for those with the poorest outcomes will help, especially in the UK where net job losses for those with poor education are greater than the EU15 average but perhaps even more important will be to ensure that productivity enhancing technologies lead to improved living standards, including lower prices for both consumers and businesses and stronger real wage growth which in turn will help drive overall job creation.